The rules they are a changin’
For many decades, the corporate world has obeyed a basic rule of secrecy. Without telling anyone, you hired the right people, provided them with the necessary means, and after working in secret, thus giving you an advantage over your competitors, you put it on the market. The surprise factor, the need to catch your competitors unprepared and so position yourself as a pioneer, was an absolutely fundamental element in business.
Any leaks by employees that could give a clue to the competition was seen as disloyalty. The etymology of the word strategy, which, like many other business terms, such as “campaign”, “target” or “VUCA”, comes from the military:
The word strategy comes from the Greek στρατηγία (strategia = art of directing armies). Στρατηγία is made up of στρατός (stratós = army), ἄγω (ago = I lead, I conduct)
No general in their right mind would tell the enemy how he intends to direct his army in a battle. Secrecy regarding tactics and movements was logical within the business environment. And in that environment, one company stood out above the others: Apple. Apple’s plans were always a mystery, its product launches were intriguing and then surprising, its strategy was speculated on… until very recently, when some analysts came up with techniques such as looking at the profiles of its employees or closely monitoring the suppliers of its value chain, it was normal an Apple event to be surrounded by a halo of mystery, conjecture and expectation.
But in the machine learning environment, this characteristic of Apple’s began to be a problem. A company that had never had problems attracting or retaining talent, began to see that people turned down its job offers. On the one hand, its obsession with user privacy made it extremely difficult to obtain data with which to feed algorithms. On the other hand, a vocation of total secrecy prevented its staff from doing what they need to do: attend conferences, talk to other experts in a nascent field where there are few rules, or raising questions to study possible reactions. Working in a bunker can be hard work, and more so if you are working on something that requires resources beyond your own.
Aware of the problem, Apple began working on differential privacy models, somewhat different to its black and white approach, and announced its incorporation into Partnership on AI, the alliance that includes Amazon, IBM, Facebook , Google and Microsoft, aimed at sharing advances and best practices in the machine learning and artificial intelligence fields. Now, it has taken another step: the company has launched a website, Apple Machine Learning Journal, with the idea of allowing its experts to tell the world what they are doing and also to attract talent to the company.
There is no comments thread, and its first article, “Improving the realism of synthetic images”, is basically an academic paper already published in arXiv in December of last year, giving an idea of the work carried out by the company. Nevertheless, it is a way to allow these experts to submit their work to ideas and input from others working on similar topics in a complex new discipline, in which collaboration is essential. At the same time, it is a way of showcasing its developments in a non-academic publication, which are generally only seen by highly specialized audiences. It is also a way of showing that something is changing in the company, in the hope this will attract people working in this area to come and work for it. The page notes: “If you’re machine learning researcher or student, we’d love to hear your feedback and questions.”
Secrecy may be important in certain disciplines and activities. But the days of James Bond style managers, obsessive NDAs and the “if I told you I would have to kill you” mentality are changing. Companies that do not provide at the least a glimpse what they are doing, which locks itself away and is surrounded by a halo of mystery, risks losing its ability to lead in a largely unexplored terrain where the incorporation and retention of talent are crucial.
For Apple, no doubt, a change in its communication style that brings it a little more into line with the approach of companies like Google or Facebook: no doubt, a smart move. And for others, a sign that, in communication, times are changing.
(En español, aquí)
